Both assessment developers and users share the responsibility for conducting some validation work with instruments they design or choose to use, before they actually use them. Validation is necessary to support particular interpretations of an assessment’s results in given populations and decision-making contexts .
The five major types of
evidence that we could use to get affirmation of validity are:
- evidence of content-based
validity (content
relevance and content
representativeness),
- evidence of validity
in response/scoring processes,
- evidence of validity
of internal structure,
- evidence of convergent
and discriminant
validity (relationships with other variables), and
- evidence of criterion-related validity (concurrent validity and predictive validity)
In particular applications, we might need only one, some, or several kind(s) of evidence to defend the assessment use. For this reason, validation studies vary in formality and extensiveness.
Page 4